Despite the ubiquity of between-talker differences in accent and dialect, little is known about how listeners accommodate this source of variability in online language comprehension. Here we sought to identify constraints on this process, in order to inform candidate theories. Three experiments used the visual world paradigm to examine the roles of memory and contextual cues to talker identity in the accommodation process. Listeners interpreted the speech of a male talker with an unfamiliar regional dialect of American English, in which the /æ/ vowel is raised to /eɪ/ only before /g/ (e.g., bag is pronounced /beɪg/), and a female talker without the dialect. We examined interpretation of words like back in the context of a competitor that has the same vowel in the familiar dialect only, as well as words like bake, which share a vowel with the competitor (bag) in the unfamiliar dialect only. In all three experiments, listeners rapidly used their knowledge of how the talker would have pronounced bag to either rule out or include bag as a temporary cohort competitor, in a talker-specific manner. Even though talkers randomly alternated across trials, providing an early cue to talker identity in the form of a preamble (Exp.1) or a portrait (Exp.2) did not overwhelmingly improve performance compared to performance in the absence of a cue (Exp.3). These results suggest that talker adaptation is rapid, even in multi-talker contexts, and that on-line adaptation processes access and use information learned during previous experiences with a talker based on minimal acoustic information.
PDF
download
878987797
028-85220240
